re: When we load an app onto a device for debugging, we lock the database so
it won't move and affect the mapping of memory addresses to functions.

Makes sense ... and, I lock the chunk I need to modify.

re: We also have the DB open for writing, since breakpoints are implemented
by changing the object code in memory -- a breakpoint modifies the
instructions so when they're executed the debugger is activated.

Understood.  But, this works fine Debugging with POSE 3.4 (and others).  It
only fails with PalmSim.  There must be something different about the
technique being used on the two platforms.  Is there something unique about
getting access to the .prc on PACE/5.0?  If not, the DB should only need to
be open for writing during the breakpoint "install".

Regards,

Terry



--
For information on using the Palm Developer Forums, or to unsubscribe, please
see http://www.palmos.com/dev/tech/support/forums/

Harry,

> Does someone have a short snippet of code that pops up an alert box as a
> notification that allows the user to cancel, but that disappears once the
> work it's alerting the user to has competed.
>
> I'm thinking of something similar to the pop-ups that appear when the
network
> code is dialling your ISP.  It says nice things like "Dialling xyz", then
> something like "Authenticating" etc.
>

Look up the Progress Manager in your documentation, it displays a progress
dialog as you want.

Christian
Falch.net DeveloperStudio for Palm OS
<http://falch.net/>


--
For information on using the Palm Developer Forums, or to unsubscribe, please
see http://www.palmos.com/dev/tech/support/forums/

Does someone have a short snippet of code that pops up an alert box as a
notification that allows the user to cancel, but that disappears once the
work it's alerting the user to has competed.

I'm thinking of something similar to the pop-ups that appear when the network
code is dialling your ISP.  It says nice things like "Dialling xyz", then
something like "Authenticating" etc.

Thanks in advance,

H.

--
For information on using the Palm Developer Forums, or to unsubscribe, please
see http://www.palmos.com/dev/tech/support/forums/

Hi everyone,

I really need your help urgently!!!

Could anyone help or advise me what to use for tapping and drawing the
gadget? I am quite new to using the gadget. I am suppose to write a
application that which is quite similar to the week schedule of the
datebook. Generally my screen will just display the table where the rows are
the times(8:00am - 6:00pm) and the columns are the (M - S).

If the user taps on the any appointment bar of the table, it will go to
another form.
If the user taps anywhere except outside the borders of the gadget and the
appointment bars, it will go to another different form.

I have taken a look at the source code of the week view of the datebook and
I spent quite some time looking but get nowhere.

What things do I have to take note?
What are the APIs function I got to use?

Any comments are welcomed. Thanks in advance. Please help me.

Low Pui Kuen

_________________________________________________________________
MSN Photos is the easiest way to share and print your photos:
http://photos.msn.com/support/worldwide.aspx


--
For information on using the Palm Developer Forums, or to unsubscribe, please
see http://www.palmos.com/dev/tech/support/forums/

Hi everyone,

     I have a gadget and 4 grouped push buttons. Why is it that whenever
I insert the pendownEvent, I could not get a respond from the either 4 of
the buttons i.e when the selected button is not highlighted?

     However when I removed the pendownEvent code, the grouped buttons works
as it should. I did not add in any code in the pendownEvent.

>>case pendownEvent :

>>handled = true; break;

Thanks in advance
Low Pui Kuen


_________________________________________________________________
MSN Photos is the easiest way to share and print your photos:
http://photos.msn.com/support/worldwide.aspx


--
For information on using the Palm Developer Forums, or to unsubscribe, please
see http://www.palmos.com/dev/tech/support/forums/

On Sat, Feb 23, 2002 at 02:50:22PM +0300, Alexander Horoshilov wrote:
> >
> > Let's say p is a pointer to a _record. The compiler requires knowledge
> > of the size of the object pointed to by p (in order to support
> > operations such as p++, or declaring an instance of _record on the
> > stack, etc.). It does not matter whether or not you actually perform
> > such an operation anywhere in the code, the compiler still needs to know
> > the size of _record. Since BitmapType is a incomplete structure definition,
> > the size of _record is unknown, and so the compiler is warning you about
that.
>
> No, standard C++ compiler *should not* know the size of the object or
> the structure in order to declare the pointer to its instance, only its
> *existence*. For example, following code should be compiled by any
> ANSI/ISO C++ compiler (actually, I don't know if CW compiles this, since
> I don't use it):

Sure, and this works also in C, not only C++.
Declaring a pointer to an unknown structure is a technique fundamental
to information hiding in C.
Too often poor coded interfaces expose data structure when not
necessary (perhaps because of the ignorance of this simple language
feature).

Ciao,
Marco

--
========================================================================
Marco Pantaleoni                                  panta@...
Padova, Italy                                              panta@...
elastiC language developer                   http://www.elasticworld.org

--
For information on using the Palm Developer Forums, or to unsubscribe, please
see http://www.palmos.com/dev/tech/support/forums/

Hello Jim,

I was unable to prevent myself from this remark :)

Jim Schram wrote:
>
> At 10:54 AM -0800 2002/02/22, Dave Carrigan wrote:
> >Jim Schram <schram@...> writes:
> >
> >> Your code *does* declare a structure containing a BitmapType
> > > structure, who's individual field definitions are not known without
> > > #defining ALLOW_ACCESS_TO_INTERNALS_OF_BITMAPS. The compiler needs to
> >> know the BitmapType structure's size information, to compute the size
> >> of a _record, and thus ultimately the size of a GraphicsRecord class.
> >
> >No, it *does not*, it declares a *pointer* to a structure containing a
> >BitmapType structure.
>
> Sorry Dave, but you still seem to be missing the very point the compiler is
alerting you to.
>
> Let's say p is a pointer to a _record. The compiler requires knowledge
> of the size of the object pointed to by p (in order to support
> operations such as p++, or declaring an instance of _record on the
> stack, etc.). It does not matter whether or not you actually perform
> such an operation anywhere in the code, the compiler still needs to know
> the size of _record. Since BitmapType is a incomplete structure definition,
> the size of _record is unknown, and so the compiler is warning you about that.

No, standard C++ compiler *should not* know the size of the object or
the structure in order to declare the pointer to its instance, only its
*existence*. For example, following code should be compiled by any
ANSI/ISO C++ compiler (actually, I don't know if CW compiles this, since
I don't use it):

struct _record;

class GraphicsRecord {
private:
   const _record * data;

The knowledge about size of the object is needed *only* when one really
tries to perform some operation on it: create new instance, declare
static or stack instance, declare new container type for it etc. What
Dave does, it *declares* the container - struct _record. Whenever
compiler finds declaration of new type:

    const struct _record {
      UInt16 flags;
      BitmapType bitmap;
    };

it required that all contained typed have to be *completely* defined at
that point:

> It *does not matter* if you're only declaring/using a pointer to such
> a structure... the fact that you are declaring a structure which contains,
> as a member field, a BitmapType structure is the reason your code fails
> without ALLOW_ACCESS_TO_INTERNALS_OF_BITMAPS defined. I'm not sure how
> much clearer I can explain this. It's Ben's point and mine.

Yes! It's the reason of the warning. The declaration of the pointer
seems to have nothing with it.

My suggestion is to put forward declaration in header and actual
declaration in CPP file and compile the latest with #define
ALLOW_ACCESS_TO_INTERNALS_OF_BITMAPS.

> >I was using a perfectly legitimate technique
> >that's been broken by Palm's attempt to enforce data abstraction beyond
> >what C can really provide. I'm not saying that the attempt was a bad
> >idea, but it will break some legitimate code in the process. However, as
> >I said before, I can work around that at the expense of making my code a
> >little less clear (and already have, in fact).
>
> I disagree that it's a "perfectly legitimate technique" but that is
unimportant.
> We agree on the result, but disagree on the rationality behind it(?) Fine...
> whatever... you've worked around the problem, so this discussion has degraded
> into little more than an exercise in C linguistics.
>
> Best Regards,
>
> Jim Schram
> PalmSource Inc.
> Partner Engineering

WBR,
--
Alexander Horoshilov

--
For information on using the Palm Developer Forums, or to unsubscribe, please
see http://www.palmos.com/dev/tech/support/forums/

> After much interaction with a 5.0 developer (name withheld to protect
the
> innocent), we determined my application actually runs absolutely
wonderful
> on PalmSim and 5.0.  As long as the Debugger isn't involved, that is.
Which
> took quite a while to figure out, because I was the only way I was
testing
> my app was while running under PalmSim and 5.0 from the debugger so I
could
> trap "surprises".
>
> So, my application that runs fine from the Debugger with POSE 3.4 (and
any
> version of Palm OS up to 4.1) has a problem with the Debugger and
> PalmSim/5.0.  It appears that the Debugger is opening the .prc with
Write
> access, and leaving it that way for the duration of the debug session.
So,
> my application code listed above trips on the "DmOpenDatabase"
function.
>
> My very patient 5.0 developer contact suggested I contact MW, and in
> particular, Ben Combee, via this newsgroup to report the ... situation
??

I'll forward this on to our debugger engineer, but I don't think there
is a good fix.  When we load an app onto a device for debugging, we lock
the database so it won't move and affect the mapping of memory addresses
to functions.  We also have the DB open for writing, since breakpoints
are implemented by changing the object code in memory -- a breakpoint
modifies the instructions so when they're executed the debugger is
activated.



--
For information on using the Palm Developer Forums, or to unsubscribe, please
see http://www.palmos.com/dev/tech/support/forums/

hi,
i am working on palm development. i m doing a project
on localisation of palm.
i need to develope hindi graffiti.
i guess, i ll be needing the graffati manager for the
same.
i m using gnu prc-tools 2.1 and sdk 4.0
my current status is that i can create glyphs.
i also have the graffiti manager api reference.
but since i dont have the programming steps i cant go
ahead.
i dont undersand the meaning of various parameters
like flagptr,uncertain length, data length,etc.i also
dont know what is a graffiti point buffer.
pls can somebody help me with it?
pls email me at picasso0072001@...
bye,
siddharth saxena

________________________________________________________________________
Looking for a job?  Visit Yahoo! India Careers
       Visit http://in.careers.yahoo.com

--
For information on using the Palm Developer Forums, or to unsubscribe, please
see http://www.palmos.com/dev/tech/support/forums/

> Then call it abstraction. You know how I'm going to react to the
> word "hide" anyway ;-)

Another term that is often used is "information hiding", which is
somewhat different from abstraction. When the term "hide" is used in
connection with programming, you should almost always assume that is
what is meant.

Cheers,
Jim Cooper

_____________________________________________

Jim Cooper      jcooper@...
Tabdee Ltd      http://www.tabdee.ltd.uk

TurboSync - Writing Palm conduits with Delphi
_____________________________________________

--
For information on using the Palm Developer Forums, or to unsubscribe, please
see http://www.palmos.com/dev/tech/support/forums/

On Fri, 22 Feb 2002, Chris DiPierro wrote:
> Hrm, the docs look old on the site. I know I documented my changes, but
> they're not up there :P

   :) busy busy busy :P

   send me another diff.. and, i'll put it in.. i believe i also changed
   some of your code to get it working under win32 (it worked fine under
   solaris) :P

// az
aaron@...
http://www.ardiri.com/


--
For information on using the Palm Developer Forums, or to unsubscribe, please
see http://www.palmos.com/dev/tech/support/forums/

Hello,

I'm testing our apps on POSER 3.4 against multiple device ROMs all which
report running sysFtrNumROMVersion = 0x03523000, specifically Visor Prism &
SPT1700.

     FrmHandleEvent <frmCloseEvent>
     FrmHandleEvent <winExitEvent>

    When closing a modal form, SPT1700 is happy, Visor Prism ROM causes 'read
from unallocated chunk of memory'.  Does this mean not only do we have to
keep up with the constantly changing PalmOS versions generating different
events, but now we have to check licensees within the same versions too?

This can not be the case, we must be doing something else wrong....

Thanks for your help,
George Aslanis
Omni Software Systems, Inc.


--
For information on using the Palm Developer Forums, or to unsubscribe, please
see http://www.palmos.com/dev/tech/support/forums/

I posted the following message on 2/21/2002 (subject: PalmSim / 5.0 ??)

**** START REPOST

This code has been running perfectly fine on Palm OS 3.1 through 4.0 in POSE
3.4.

<snip>
  // Open the resource database for read/write
  tmpDBH = DmOpenDatabase(cardNo, dbID, dmModeReadWrite);
  if (tmpDBH == NULL)
   {
   err = DmGetLastErr();
   AbandonUpdates();
   return (err);
   }
<snip>

In PalmSim with Palm OS 5.0, I get an error code 534
(dmErrAlreadyOpenForWrites).

OK, so what's unusual about this ??

Well, I'm updating my own .prc.  Yes, I need to.

**** END REPOST

After much interaction with a 5.0 developer (name withheld to protect the
innocent), we determined my application actually runs absolutely wonderful
on PalmSim and 5.0.  As long as the Debugger isn't involved, that is.  Which
took quite a while to figure out, because I was the only way I was testing
my app was while running under PalmSim and 5.0 from the debugger so I could
trap "surprises".

So, my application that runs fine from the Debugger with POSE 3.4 (and any
version of Palm OS up to 4.1) has a problem with the Debugger and
PalmSim/5.0.  It appears that the Debugger is opening the .prc with Write
access, and leaving it that way for the duration of the debug session.  So,
my application code listed above trips on the "DmOpenDatabase" function.

My very patient 5.0 developer contact suggested I contact MW, and in
particular, Ben Combee, via this newsgroup to report the ... situation ??

Regards,

Terry



--
For information on using the Palm Developer Forums, or to unsubscribe, please
see http://www.palmos.com/dev/tech/support/forums/

> It means hiding implementation behind defined interfaces, so that it's
> hidden from code even though it may well still be in public view.  Go
> find a programming book and look up "abstraction".

	 Then call it abstraction. You know how I'm going to react to the
word "hide" anyway ;-)

> Hiding internals from applications is a step *towards* an open,
> well-documented OS.

	 We agree.



--
For information on using the Palm Developer Forums, or to unsubscribe, please
see http://www.palmos.com/dev/tech/support/forums/

> Most all major corporations throughout the world are using code without
> auditing it themselves, and this includes OSSS.

	 Most major corporations are regularly exploited externally and
internally on a regular basis because of insufficient security measures at
the application, OS, and hardware level to secure their own data. Just
because it doesn't make it to 'news.com' doesn't mean it isn't happening.
I've dealt with this dozens of times at a previous job at a very large
proprietary pharmeceutical company.

	 Nobody is advocating making the code "free", or publically
accessible. I said "open", which means access to it should be available.
Whether that access is through NDAs, or other methods that do not require
restrictive licensing, that's all that matters.

> See http://news.com.com/2100-1001-830130.html, where you will discover
> that "...the programs are getting audited a lot less than people think".

	 A moot point. Nimda is still pounding machines across the internet
every single day. I get 30-50 new hosts _IN MY SUBNET_ a week. How long ago
were people told to patch their machines against this? 7 months?

	 Auditing seems to happen when a problem is found, not when an
application is designed. This is 2002, and it's time to start thinking about
scalable, secure, distributed applications. You can't put security in a
black box anymore. Everyone's learned their lessons already by getting
burned by companies like Microsoft, Oracle, and others.

> For example, Sendmail has been open source for 20 years, and people have
> found more vulnerabilities in it every year of that 20 years.

	 Security is a process, not a program.

	 That being said, sendmail is developed by a company, with employees
and an agenda. Just because I find a hole, or an exploitable section of
code, and report it, is no guarantee that _ANYTHING_ will be done about it
(witness any of the hundreds of still-open holes in that OS from our friends
in Redmond)

	 My point is that auditing ensures data security more than "assuming"
that an application is secure. Everyone should be learning their lessons now
after getting burned by all the "spyware" in applications these days from
RealPlayer, Windows Media Player, :CueCat, and others.



/d



--
For information on using the Palm Developer Forums, or to unsubscribe, please
see http://www.palmos.com/dev/tech/support/forums/

I'm kind of glad I posted that method since it made me take another look at
it. Turns out there's a much better way to do it. Here's the new
implementation:

Err PktLibGetRecordByID(DmOpenRef dbP, UInt32 id,
     PktLibDBRecordPtr recordP, MemHandle *recordH)
{
UInt16  index = 0;

     if (DmFindRecordByID(dbP, id, &index) == 0)
         return PktLibGetRecord(dbP, index, recordP, recordH);
     else
         return 1;
}

Boy, that was kind of embarrassing. :-)

"CausticMango" <samuelwford@...> wrote in message
news:77615@palm-dev-forum...
>
> I use UniqueID to link records within the same DB and it works fine (so
> far). I use the following, admittedly inefficient, method to locate a
record
> by it's ID:
>
> /************************************************************
>  *
>  * FUNCTION:    PktLibGetRecordByID
>  *
>  * DESCRIPTION: Get a record from the Database by its UniqueID
>  *
>  * PARAMETERS:  database pointer - open db pointer
>  *              uniqueid - localid of record to lock
>  *              record pointer - pointer structure
>  *              record - MemHandle to unlock when done
>  *
>  * RETURNS:     ##0 if successful, errorcode if not
>  *
>  * The record's MemHandle is locked so that the pointer to
>  * strings within the record remain pointing to valid chunk
>  * versus the record randomly moving.  Unlock the MemHandle when
>  * PktLibDBRecord is destroyed.
>  *
>  *************************************************************/
> Err PktLibGetRecordByID(DmOpenRef dbP, UInt32 id,
>     PktLibDBRecordPtr recordP, MemHandle *recordH)
> {
> UInt16  index = 0;
> UInt16  num, attr;
> UInt32  rid;
>
>     num = DmNumRecords(dbP);
>     while (index < num) {
>         DmRecordInfo(dbP, index, &attr, &rid, NULL);
>         if (id == rid && !(attr & dmRecAttrDelete))
>             break;
>         index++;
>     }
>
>     if (index < num)
>         return PktLibGetRecord(dbP, index, recordP, recordH);
>     else
>         return 1;
> }
>
> It MAY be possible to use the memory manager to get a direct handle to the
> record by ID using MemLocalIDToLockedPtr, but I'm not sure, I haven't
tried
> it out.
>
> "Ben A. HiLLeLi" <bahillel@...> wrote in message
> news:77587@palm-dev-forum...
> >
> > After too much deliberation I've decided on using the PalmOS assigned
> > UniqueID as a "primary key" and as the lookup attribute that will "link"
> two
> > related databases:
> > i.e. AllergyRecord.PatientIDLink=Patient.UniqueID...
> >
> > to get patient referenced in AllergyRecord:
> > DmFindRecordByID(gPatientDB, AllergyRecord.PatientIDLink, &index_out)
> > DmQueryRecord(gPatientDB, index_out)
> > or something like that...
> >
> > I have not invested that much time in the conduit yet and want to know
if
> > there is sample code on importing record UniqueID's and/or the entire
> record
> > headers and using these as attributes in a desktop database... I think
> once
> > I get the UniqueID onto the PC (and into PC Endian) it won't be a
problem
> > assigning as a value to a dbase attribute...
> >
> > SideQ: When working with alphabetically sorted databases, what other
> > efficient methods are there of doing a quick key lookup. Binary Searches
> > would seem rather inefficient for they would require resorting by key,
> then
> > sorting back alphabetically. I'm looking for a backup (insurance :-O) to
> the
> > above method as I've heard experienced Palm developers say "USE the
> > uniqueID, works for me" and "Don't Use uniqueID, it can change and/or
not
> be
> > unique"... I was thinking of a two part key, one part Char: the letter
the
> > name starts with (used only for quicker searches as names can change)
and
> a
> > permanent ID (iterative, last used stored in preferences)... Hopefully
the
> > first method workz (no insurance premiums...missed deadlines)
> >
> > Ben
> > p.s. from what I've read in the documentation (briefly skimmed),
> UniqueID's
> > should stay constant during synchronization using more current versions
of
> > HotSynch... I'm a beginner looking for answers though, do NOT take my
word
> > on it AND/OR/PLEASE correct me...
> >
> >
> >
> >
> >
> >
>
>
>
>



--
For information on using the Palm Developer Forums, or to unsubscribe, please
see http://www.palmos.com/dev/tech/support/forums/

Thanks Joe.

Just what I was looking for. I should say that I am not only new to
programming for palm, but also in C. Thanks for your help. That was very
useful!

Thanks again.

Scott Mebberson


-----Original Message-----
From: bounce-palm-dev-forum-46917@...
[mailto:bounce-palm-dev-forum-46917@...] On Behalf Of Joe
Sent: Friday, 22 February 2002 5:17 PM
To: Palm Developer Forum
Subject: Re: Warning messages?

--- Scott Mebberson wrote:
>
> frmProjectMain.c:64: warning: assignment makes integer from pointer
> without a cast
> ...
> the line is: GlobalPrjct = num;
>
> I have defined GlobalPrjct as UInt16 GlobalPrjct; and num as static
> Char num[26];
>
> I am totally new to this, so please go easy on me.
>

Darn!  What's the fun in that?

Do you understand what data types are?

You declared num to be an array of 26 Chars.  In C, arrays and pointers
are very similar.  You can think of num as a pointer to the first
character of that array of 26 Chars.  So, num is the pointer that
Falch.net is complaining about.

You declared GlobalPrjct as an unsigned 16-bit integer.

Then you tried to assign the pointer to the integer and your compiler
warned you that that is a no-no.

Why are you trying to assign the string (num) to GlobalPrjct?  If you
want to save the string in a global, do it like this:

static Char someGlobal[26];
...
StrCopy(someGlobal, num);

Hope that helps.


__________________________________________________
Do You Yahoo!?
Yahoo! Sports - Coverage of the 2002 Olympic Games
http://sports.yahoo.com

--
For information on using the Palm Developer Forums, or to unsubscribe,
please see http://www.palmos.com/dev/tech/support/forums/


--
For information on using the Palm Developer Forums, or to unsubscribe, please
see http://www.palmos.com/dev/tech/support/forums/

> At 2:03 PM -0800 2/18/02, Keith Rollin wrote:
> >Perhaps there's a problem with the way I convert ticks to
> >microseconds?  Here's my code:
> >
> > // Convert ticks to microseconds.
> >
> > hostTimeout.tv_sec = timeout / sysTicksPerSecond;
> > hostTimeout.tv_usec = (timeout % sysTicksPerSecond) *
> >(1000000 / sysTicksPerSecond);
> > hTp = &hostTimeout;

--- Today, Keith Rollin wrote:
> What?  No one found the problem?  So much for peer code reviews....
> :-)

Hey, I thought you were being rhetorical!


__________________________________________________
Do You Yahoo!?
Yahoo! Sports - Coverage of the 2002 Olympic Games
http://sports.yahoo.com

--
For information on using the Palm Developer Forums, or to unsubscribe, please
see http://www.palmos.com/dev/tech/support/forums/

--- "David A. Desrosiers" wrote:
> I for one would not be using any code that
> purports to be "secure" without auditing it myself,
> nor would any major corporation worth their salt and
> who valued their data.

AIA for the OT remark:

This is classic hyperbole and wishful thinking, probably by an open
source afficionado.  Most all major corporations throughout the world
are using code without auditing it themselves, and this includes OSSS.
See http://news.com.com/2100-1001-830130.html, where you will discover
that "...the programs are getting audited a lot less than people
think".

The "peer review" argument from open-source advocates has merit, but
it's not a guarantee of secure software. For example, Sendmail has been
open source for 20 years, and people have found more vulnerabilities in
it every year of that 20 years.


__________________________________________________
Do You Yahoo!?
Yahoo! Sports - Coverage of the 2002 Olympic Games
http://sports.yahoo.com

--
For information on using the Palm Developer Forums, or to unsubscribe, please
see http://www.palmos.com/dev/tech/support/forums/

On Fri, Feb 22, 2002 at 03:57:10PM -0800, David A. Desrosiers wrote:
>> With Palm OS 5, remember that the decision to hide internals is a
>> technical one rather than some new philosophy.
>
>  Can you go into a bit more depth of the term "hide"? What exactly
> are you "hiding", and what is the technical hurdle that necessitated closing
> it off from public view?

Good God.  "Hide" here doesn't mean closing things off from public view.
It means hiding implementation behind defined interfaces, so that it's
hidden from code even though it may well still be in public view.  Go
find a programming book and look up "abstraction".

>  Actually, having an open, well-documented OS will extend it more

Hiding internals from applications is a step *towards* an open,
well-documented OS.

     John  "there is no conspiracy :-)"

--
For information on using the Palm Developer Forums, or to unsubscribe, please
see http://www.palmos.com/dev/tech/support/forums/

I have a legacy app that uses a special font for some of its
controls. Back then, I implemented this by defining the resource as
having font 0, then in the app I would do:

    // At startup:
    FntDefineFont(192, MemHandleLock(DmGetResource('NFNT', 192)));

    // During frmOpenEvent:
    ControlType* ctl = FrmGetObjectPtr(...);
    ctl->font = 192;

This seemed to work fine.

Fast forward to today, and I have another app that uses a special font
in some of its controls. I implemented it this time by defining the
resource as having font 192, and calling FntDefineFont() before any
forms are loaded. This also works fine, on PalmOS 3.2 and above. On 3.0
and 3.1 the controls continue to use font 0. So, I decided to use the
same method as the legacy app, and changed my resource to use font 0,
and changed my code to do

    // During frmOpenEvent:
    ControlType* ctl = FrmGetObjectPtr(...);
    CtlGlueSetFont(ctl, 192);

this also works fine on 3.2 and above, but still doesn't work on 3.0 and
3.1. Further experimenting showed that I could make it work with
CtlGlueSetFont(ctl, boldFont) and other system fonts, but wouldn't work
with any of my installed fonts.

Now here is where it gets really weird. I put both my legacy app and the
new app on pose running a 3.0 rom. After running the legacy app and
verifying it was using special fonts, I loaded the new app, and much to
my surprise, it was now working correctly. I did a reset, and the new app
reverted to the incorrect behavior. I ran the legacy app again, went
back to the new app, and it was working again, and it continues to work
fine until I reset the device. After a reset, it won't work until I
first run my legacy app. I don't even have to load a form in the legacy
app that uses a custom font; simply loading the app seems sufficient. It
also doesn't seem to make a difference whether I use the
define-the-font-in-the-resource method or I use the CtlGlueSetFont
method in the new app.

I'm pretty sure that it's not a problem with my application, since it
works fine on any ROM from 3.2 up to 4.1, and also works fine in the
simulator. There seems to be a bug with PalmOS 3.{0,1}, and my legacy
app is inadvertently triggering something that installs a work around,
at least until the Palm resets.

Any ideas?

--
Dave Carrigan (dave@...)            | Yow! I'm not available for
UNIX-Apache-Perl-Linux-Firewalls-LDAP-C-DNS | comment..
Seattle, WA, USA                            |
http://www.rudedog.org/                     |

--
For information on using the Palm Developer Forums, or to unsubscribe, please
see http://www.palmos.com/dev/tech/support/forums/

At 11:11 AM -0500 2/18/02, Matt Hebley wrote:
>At 02:33 PM 15/2/2002 -0800, Keith Rollin wrote:
>>At 3:56 PM -0500 2/15/02, Matt Hebley wrote:
>>>From what I understand, POSE using the host TCP stack has a
>>>hardwired 2 second timeout (i.e. it ignores the timeout you
>>>specify in the NetLib call).
>>
>>From where do you understand this?  Poser does not ignore the
>>timeout values and instead substitute a 2 second timeout.  There's
>>nothing in the documentation, release notes, or source code that
>>says anything about 2 second timeouts, so I'm keen to learn where
>>you heard this.
>
>I understand this mainly from hearsay, and personal experience. In
>Greg Winton, Palm OS Network Programming, O'Reilly, p199, it says
>"On Windows and Max machines, the host's sockets API do not support
>timeouts, so this parameter is ignored." It then goes on to say that
>POSE appears frozen until the call "times out." I asked at
>PalmSource and nobody seemed to know too much about it, except that
>there used to be a bug that had been fixed.

Perhaps the two-second timeout theory comes from AppNetTimeout, which
is defined to be two seconds by default.  However, that's a
NetLib/sockets compatibility value, and not something directly
managed by Poser.

So, to summarize:

* Two-second timeouts are probably due to AppNetTimeout, and not
anything hardwired in Poser.

* Comments about Poser ignoring timeout values are no longer true
with regards to sending and receiving data (as of Poser 3.1).  Other
NetLib functions taking timeouts and corresponding to sockets
functions that don't take timeouts (such as NetLibSocketOpen or
NetLibSocketClose) still ignore timeouts.

* Poser's timeouts will be off by a factor of 10 on Windows and Unix,
and a factor of .6 on the Mac.  This will be fixed in Poser 3.5.

* Poser will still be unresponsive while redirected NetLib calls timeout.

--

-- Keith Rollin
-- Palm OS Emulator engineer

--
For information on using the Palm Developer Forums, or to unsubscribe, please
see http://www.palmos.com/dev/tech/support/forums/

What?  No one found the problem?  So much for peer code reviews.... :-)

Steve Lemke provided me the hint.  The problem is with
"sysTicksPerSecond".  This is defined to be 100 when building a Palm
OS ROM, but 1000 when building Windows applications (and 60 when
building Macintosh applications).  So there's your factor of 10...

I'm pretty sure I can fix this for Poser 3.5.

-- Keith Rollin
-- Palm OS Emulator engineer


At 2:03 PM -0800 2/18/02, Keith Rollin wrote:
>At 11:11 AM -0500 2/18/02, Matt Hebley wrote:
>>In my code, I set the timeout to 20 seconds which appears to be the
>>case on a real device (I haven't actually set up a scenario that
>>will take forever and timed the timeout). With POSE the same code
>>often times out on the host stack in what I estimate to be 2
>>seconds (definitely way less than 20 seconds), but doesn't seem to
>>do the same using the NetLib stack.
>
>Perhaps there's a problem with the way I convert ticks to
>microseconds?  Here's my code:
>
> // Convert ticks to microseconds.
>
> hostTimeout.tv_sec = timeout / sysTicksPerSecond;
> hostTimeout.tv_usec = (timeout % sysTicksPerSecond) *
>(1000000 / sysTicksPerSecond);
> hTp = &hostTimeout;


--
For information on using the Palm Developer Forums, or to unsubscribe, please
see http://www.palmos.com/dev/tech/support/forums/

> I have been trying for some time to get Palm and PC communicating through
IrDA.

This really belongs on comm-dev-forum.

Which version of Palm OS and which version of Windows?
Which IrDA protocol do you want to use?

> On the PC side I use the following code (I am using VC 6.0):

I can't help with this part - I don't do windows. :)

> On the Palm side I have:
>
> IrOpen
> IrBind
> IrSetConTypeLMP
> IrSetDeviceInfo
> IrDiscoverReq
> IrDataReq

So I guess you want to use IrLMP. Be sure you set up the Windows side to use
the same protocol. You need to do a connect at the LAP level and then a
connect at the LMP level after you discover and before you send data. Also,
you can't just make these OS calls sequentially. You need a state machine so
that you can proceed with each step only after the previous step completes.
You're notified when an action completes, and whether it succeeded or
failed, via a callback. You should retry at least a couple of times if a
call fails, either immediately or after a while.

> Any sample code or suggestions are welcome.

The canonical example is Till Harbaum's IR Ping:
   http://bodotill.suburbia.com.au/ir_ping/ir_ping.html
It lets you execute each step one at a time, letting you be the state
machine yourself.

BeamBooks source is now public domain:
   http://www.appliedthought.com/beambooks/
You can download a StuffIt archive of a Mac CodeWarrior project. Use StuffIt
Expander for Windows:
   http://www.aladdinsys.com/
to unstuff it on your PC and then look at IrConnection.c. It provides a
layer over IrLMP that lets you send arbitrary chunks of data back and forth
between two devices. It starts with a symmetrical connection: both sides
must initiate.
--
Danny @ PalmSource

--
For information on using the Palm Developer Forums, or to unsubscribe, please
see http://www.palmos.com/dev/tech/support/forums/

I'm really not the right person to answer your questions, and I'm sure I'll
make some mistakes, but here goes anyway...

The decision to disallow access to internals of some structures was made
because those structures have a different layout in memory on an ARM than
they do on a 68K. The differences are caused by two things: endianness and
packing rules. The result is that in order to expose such a structure, the
operating system would need to maintain a shodow structure that is arranged
as the 68K would expect it to be, and deal with propagating changes between
these two structures whenever either one was changed. Such shadow structures
are used in PalmOS 5, but adding a bunch more is best avoided.

That is what little I think I might know ;-) Hopefully others will correct
me if I'm wrong.
--
Peter Epstein

--
For information on using the Palm Developer Forums, or to unsubscribe, please
see http://www.palmos.com/dev/tech/support/forums/

> With Palm OS 5, remember that the decision to hide internals is a
> technical one rather than some new philosophy.

	 Can you go into a bit more depth of the term "hide"? What exactly
are you "hiding", and what is the technical hurdle that necessitated closing
it off from public view?

> In the end, I hope that it helps make Palm OS more of a real OS, but I
> don't think that's the driving force behind it.

	 Actually, having an open, well-documented OS will extend it more
than licensing agreements and locking everything in DLLs, because it allows
developers to continue to develop against it and support it with the wealth
of documentation you provide.

	 In fact, based on the new "philosophical" direction using RSA and
SSL, I sincerly _HOPE_ those relevant sections of code are open, or at least
solidly documented, to the point of being subject to a professional audit
for viability and security. I for one would not be using any code that
purports to be "secure" without auditing it myself, nor would any major
corporation worth their salt and who valued their data.

	 I'm intrigued, and would like to hear more.


/d



--
For information on using the Palm Developer Forums, or to unsubscribe, please
see http://www.palmos.com/dev/tech/support/forums/

I can't comment on what the future holds, as I really don't know. What I
will say is that you're quite right when you describe how the OS and
applications were originally developed. There was no wall between the two,
and the OS really was just enough to get the job done for the applications
that shipped on the ROM. This was probably the right thing to do at the
time. A huge amount of progress has been made towards turning Palm OS into a
real operating system over the years, and I too hope that the split will
lead to further improvements in this area.

With Palm OS 5, remember that the decision to hide internals is a technical
one rather than some new philosophy. In the end, I hope that it helps make
Palm OS more of a real OS, but I don't think that's the driving force behind
it.
--
Peter Epstein

--
For information on using the Palm Developer Forums, or to unsubscribe, please
see http://www.palmos.com/dev/tech/support/forums/

> Unfortunately, in one context it will be the app beaming itself, but in
> another it will be a beam station not running any of our software (for the
> near future anyway).

No problem - just dump a modified version of the app onto the beam station.
As far as it's concerned, it's just an arbitrary file. It doesn't care
whether the hidden attribute in the prc header is set or not.

> In my case, what's being beamed is a self-extracting archive, and
primarily
> from a beam station to people who are going to use the app right away.  In
> that context, auto-launch seems like a good idea from a user experience
> point of view.

Agreed. The WideRay stations at PalmSource auto-launched.
-
Danny

--
For information on using the Palm Developer Forums, or to unsubscribe, please
see http://www.palmos.com/dev/tech/support/forums/

"Terry Dennis" <tedennis@...> wrote in message
news:77653@palm-dev-forum...
>
> re: I'm still not following what it is you are actually trying to
accomplish
> ... why do you need the information about which group a control belongs
to?
> If you want to know something about the group selection, use
> FrmGetControlGroupSelection().  Why are you starting with the control and
> working your way up to the group?
>
> Really, it doesn't matter WHY he needs the information.  He needs it.
> That's all that matters.

Call it curiosity ... it's not like I'm holding out on a secret technique
... and
I am one of the first to fault Palm on their lack of complete APIs.

...
> I don't want to get into a political pi$$ing match about who needs what,
or
> why they need it.  The "Zen of Palm" afficionados need to realize that
they
> provide a tool.  How that tool is used depends on the person who purchased
> the tool.

Oh, don't get me started on that one.  I have _big_ questions about
whether these folks have yet realized that they are providing an Operating
System as opposed to their own private application.  Some symptoms:
     - fatal reset alerts put up instead of returning result codes listed in
the documentation
     - a file list builder that has (undocumented) exceptions, and can not
build a
         list of all dbs of a given type. (I am certain now that this is the
reason
         that some doc readers "take over" the creator ID of docs on the
system
         ... because they can't just find all doc dbs on the basis of their
type)
     - adding a ":" to whatever the programmer sets into a label in a table
(if
         I wanted a colon, I'd put one there!)
     - incomplete APIs _combined_with_ a "hands off" attitude to the relevant
         data structures
     - inconsistant treatment of form objects (you can hide _this_ but not
_that_)

The _OS_ level APIs should be application-agnostic, and give tools to the
developers.  It appears that the guys writing the PalmOS applications
(launcher,
datebook, etc) and the guys writing the OS services were not keeping a
wall between their efforts.  ("Hey, Jack ... if you'd pop up an alert in
this
system API then I wouldn't have to write code to check for errors every time
I called it."  "Ok, Fred.")

I sincerely hope that, now that they are their own company,  these guys take
a _good_ look at what their responsibilities are as an OPERATING SYSTEM
provider and take a more serious attitude to these types of problems.

....
>
> I understand the need to "hide" internals from programmers.  In fact, I
was
> surprised that the platform was so immature in the early days that it
> exposed a lot of the internals to the outside world.  Thankfully, the
> platform has matured a lot since then, and 5.0 is a big step forward.
But,
> there are still omissions which keep the user community from using the
tool
> to its potential.
>
...


--
-Richard M. Hartman
hartman@...

186,000 mi/sec: not just a good idea, it's the LAW!



--
For information on using the Palm Developer Forums, or to unsubscribe, please
see http://www.palmos.com/dev/tech/support/forums/

"Danny Epstein" <Danny.Epstein@...> wrote in message
news:77662@palm-dev-forum...

> Mark A. Peters wrote:
> > I wrote code to beam an application, and figured that simply setting the
> > goToCreator field to the app's creator ID in the Exchange Manager socket
> > for the beam would make this happen.  I was wrong though - the
application was
> > never sent the goto launch code.

> Which app is receiving the beam? Since your app isn't on the receiving
> device, it must be Launcher. [...]

Ah, of course, that makes sense - I am setting the target to the launcher.

> > I don't see a way from the application's point of view that it can tell
> > the difference of being beamed or being installed.

> This is the key. If your app is doing the beaming, then it can change the
> app "on the way out". [...]

Unfortunately, in one context it will be the app beaming itself, but in
another it will be a beam station not running any of our software (for the
near future anyway).

> This is all great fun, but you should ask yourself whether you want to
> auto-launch your app after it's beamed. [...]

In my case, what's being beamed is a self-extracting archive, and primarily
from a beam station to people who are going to use the app right away.  In
that context, auto-launch seems like a good idea from a user experience
point of view.

Mark Peters



--
For information on using the Palm Developer Forums, or to unsubscribe, please
see http://www.palmos.com/dev/tech/support/forums/